INDEX

Explanations

difficulties or obstacles

oai_token-act-pair · gpt-3.5-turbo

terminology related to challenges, problems, and criticisms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

psc

-0.79

bos

-0.72

raph

-0.72

rete

-0.70

sten

-0.67

late

-0.65

lys

-0.64

azard

-0.62

 externalToEVAOnly

-0.62

ceans

-0.61

POSITIVE LOGITS

 incent

0.82

 imaginable

0.73

 horr

0.71

 encount

0.69

å¥³

0.68

 stru

0.66

 indu

0.66

 drawback

0.65

ieties

0.65

 attraction

0.65

Activations Density 0.212%

difficulties or obstacles

terminology related to challenges, problems, and criticisms

No Comments

No Known Activations